Amplitude says Wave ships product changes in 1-2 days with OpenAI
Original title: How Amplitude ships product improvements in 1–2 days with OpenAI
Amplitude reports a 160% lift in homepage card sends and lower GPT-6.1 Sol evaluation cost.
OpenAI published a November 12, 2026 customer story on Amplitude's product agent Wave, which Amplitude says is roughly 80% powered by OpenAI models. Wave uses product data, session replays, feedback and code to propose changes, hand plans to coding agents including Codex, and measure results. Amplitude estimates shipping time of 1-2 days, down from about a month, and reports a 160% lift in homepage suggestion-card sends versus control after one day. Model scores and costs are Amplitude's own evaluations.
Unconfirmed: Engagement, cost and quality figures are Amplitude's own measurements, not independent results.
- In 100 homepage replays, 80 users typed questions and 12 selected a suggestion.
- GPT-6.1 Sol scored 4.06 of 5 on Opportunity Investigator, with a 54-minute median.
- Data Science Agent mean cost was $1.28 for GPT-6.1 Sol versus $2.63 for another model.
- GPT-6 Luna handles frequent lower-cost checks; GPT-Image-2.5 Flare makes mockups.
- Amplitude says teams review plans before coding agents implement them.
From the original article
7 min read at openai.comBuilt primarily on OpenAI models, Amplitude’s proactive product agent, Wave, identifies what to build, helps teams ship changes with coding agents, and measures the impact.
openai.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.